Disney researchers released an algorithm that automates neural face swapping for images and videos at megapixel resolutions. To accomplish this, they trained a multiway comb neural network, created a light/contrast preserving blending method, and achieved temporal stability for high resolution videos.